Can you post an example? I wrote some extensions to methylumi so that I could deal with CSV files and BeadDataPackR-compressed txt/locs files on Infinium; it would likely not be a big deal to make this work with methylumi/lumi. The latter lacked a controlData slot last time I checked, but once that's in there, it should be no big deal to go from raw data to lumi/methylumi. The defaults in GenomeStudio aren't very sensible IMHO. At some point in the past, both detection and background correction were computed by applying the averages of both channels' fluorescence for the negative control probes to the analytical probes in each channel. This doesn't make very much sense, so apparently the fix was to stop doing background correction and... I don't precisely know what happened with p-values but theirs are a little bit off versus when I compute them from scratch (and by "a little" I mean "differences show up at 20 or so decimal places"... so, not by much). The trouble with not doing background correction of any sort, however, is that the 12 slots per strip on a BeadArray assay have a smoothly increasing gradient of background intensities across the strip, with nonlinear effects on the intensity of the analytical probes in each channel (worse for green than for red, as it happens). So if you don't do background correction you end up with artifacts a considerable portion of the time. Normexp correction alone disposes of most of this effect, along with a lot of overly conservative detection p-values that result from under- or over- estimating the difference between background and signal probes. (The correlation between technical replicates nearly always improves after doing sensible background correction and recomputing the detection p-values, suggesting that this sequence makes more sense than the other way) Anyhow, it would be interesting to see what all goes into GEO and what comes out for methylation. Pan Du and the Lumi guys have automated the process of submitting a methylation dataset to GEO, hopefully with the control probe data it will be possible to include all of the data on the chip rather than just what makes it through GenomeStudio :-) On Sat, Oct 9, 2010 at 12:32 PM, S T <elengss@gmail.com> wrote: > Dear Sean, > > Thanks very much for answering my question. > . > 1. Can GEOquery be used with DNA methylation data in view of the fact that > the data itself is not just a summarised beta-value, but also includes the > p-value of detection above background, and the values in the U and M > channels. Will I be able to access the U, M and p-values from GEOquery? > . > 2. As a novice, I'm interested in using the tools in Methylumi to perform > quality control/normalization rather than writing my own script to do this > from the start etc. Hence is there any way to do this after I have imported > the data into R? > . > 3. I was told to import data by gse <- getGEO("GSExxxx") > . > then get the actual data via > data.m <- exprs(gse[[1]]), > . > and the sample information via > clin.m <- as.matrix(pData(gse[[1]]) > . > and the platform annotation matrix via > anno.m <- as.matrix(Table(gpl)) > . > Do I just type in "clin.m<- as.matrix(pData(gse[[1])"? When I type that in, > I get a + sign below that. What else do I need to type in? > . > Sorry for such a stupid question. > I tried looking elsewhere but could not find the answer. > . > ENS > > > On Sat, Oct 9, 2010 at 10:14 PM, Sean Davis <sdavis2@mail.nih.gov> wrote: > > > > > > > On Sat, Oct 9, 2010 at 10:03 AM, S T <elengss@gmail.com> wrote: > > > >> Dear all, > >> . > >> Sorry, I am a novice user of Bioconductor. > >> I saw that Methylumi can take either Final Report format or tab > delimited > >> formats. > >> May I ask whether Methylumi can import Infinium methylation data from > GEO > >> - > >> specifically GSE type files? > >> > > > > It cannot. However, once the data are loaded with getGEO(), you already > > have them in R, so there is not a need to use methylumi to load the data > > again. > > > > Sean > > > > > > [[alternative HTML version deleted]] > > _______________________________________________ > Bioconductor mailing list > Bioconductor@stat.math.ethz.ch > https://stat.ethz.ch/mailman/listinfo/bioconductor > Search the archives: > http://news.gmane.org/gmane.science.biology.informatics.conductor > -- With four parameters I can fit an elephant, and with five I can make him wiggle his trunk. John von Neumann [[alternative HTML version deleted]]